Research on the Application of SIFT Algorithm in UAV remote sensing image feature extraction

UAV remote sensing is characterized by fast speed and flexibility and applied widely. The extraction and matching for feature point of UAV image is a foundational work in the application of UAV remote sensing. SIFT is a currently widely used feature point extraction algorithm that keeps invariant as the scale changes. It is also invariant as image changes in other aspects. Meanwhile it maintains certain stability to noise and light changes. SIFT is a kind of compute-intensive mission and usually takes much time. The author made parallel processing for SIFT algorithm and good accelerator ratio has been achieved on the premise of ensuring good result of feature point extraction and matching.